MBTSAD: Mitigating Backdoors in Language Models Based on Token Splitting and Attention Distillation
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866915092096876544 |
|---|---|
| author | Ding, Yidong Niu, Jiafei Yi, Ping |
| author_facet | Ding, Yidong Niu, Jiafei Yi, Ping |
| contents | In recent years, attention-based models have excelled across various domains but remain vulnerable to backdoor attacks, often from downloading or fine-tuning on poisoned datasets. Many current methods to mitigate backdoors in NLP models rely on the pre-trained (unfine-tuned) weights, but these methods fail in scenarios where the pre-trained weights are not available. In this work, we propose MBTSAD, which can mitigate backdoors in the language model by utilizing only a small subset of clean data and does not require pre-trained weights. Specifically, MBTSAD retrains the backdoored model on a dataset generated by token splitting. Then MBTSAD leverages attention distillation, the retrained model is the teacher model, and the original backdoored model is the student model. Experimental results demonstrate that MBTSAD achieves comparable backdoor mitigation performance as the methods based on pre-trained weights while maintaining the performance on clean data. MBTSAD does not rely on pre-trained weights, enhancing its utility in scenarios where pre-trained weights are inaccessible. In addition, we simplify the min-max problem of adversarial training and visualize text representations to discover that the token splitting method in MBTSAD's first step generates Out-of-Distribution (OOD) data, leading the model to learn more generalized features and eliminate backdoor patterns. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_02754 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MBTSAD: Mitigating Backdoors in Language Models Based on Token Splitting and Attention Distillation Ding, Yidong Niu, Jiafei Yi, Ping Cryptography and Security Computation and Language In recent years, attention-based models have excelled across various domains but remain vulnerable to backdoor attacks, often from downloading or fine-tuning on poisoned datasets. Many current methods to mitigate backdoors in NLP models rely on the pre-trained (unfine-tuned) weights, but these methods fail in scenarios where the pre-trained weights are not available. In this work, we propose MBTSAD, which can mitigate backdoors in the language model by utilizing only a small subset of clean data and does not require pre-trained weights. Specifically, MBTSAD retrains the backdoored model on a dataset generated by token splitting. Then MBTSAD leverages attention distillation, the retrained model is the teacher model, and the original backdoored model is the student model. Experimental results demonstrate that MBTSAD achieves comparable backdoor mitigation performance as the methods based on pre-trained weights while maintaining the performance on clean data. MBTSAD does not rely on pre-trained weights, enhancing its utility in scenarios where pre-trained weights are inaccessible. In addition, we simplify the min-max problem of adversarial training and visualize text representations to discover that the token splitting method in MBTSAD's first step generates Out-of-Distribution (OOD) data, leading the model to learn more generalized features and eliminate backdoor patterns. |
| title | MBTSAD: Mitigating Backdoors in Language Models Based on Token Splitting and Attention Distillation |
| topic | Cryptography and Security Computation and Language |
| url | https://arxiv.org/abs/2501.02754 |